Efficient Malware Detection in 5G Environments with Deep Neural Networks
Vivek Ranjan, B.S. Babitha, Shweta Loonkar · 2024
The widespread use of Android devices on 5G networks has resulted in an increase in the number of third-party apps vying for market share. But this growth has made it an enormous task to effectively evaluate a large number of apps before launching them on the market. Due to the excessive quantity of applications, there are major bottlenecks in the current review process, which takes a lot of time and computer power. Moreover, the intricate characteristics of Android applications provide a challenge in identifying the ideal feature combination for differentiating between dangerous and benign software. Relevance is selected using the Random Forest method, and hidden patterns within features are automatically extracted using the A convolutional Neural Network (CNN). Using this combination process, the most biassed feature subset may be found. Experimental validations using real data from AMD databases and an external download platform show how effective the suggested approach is. The results show a remarkable 95.69% Fl-score index. To sum up, our strong deep learning A system, DLAMD, provides a workable answer to the problems brought on by the explosion of Android apps. In the ever-changing world of Android application development, our framework shows itself to be a dependable and effective tool for identifying software that is malicious or benign by using sophisticated feature selection algorithms and combining quick pre-detection with in-depth analysis.